Automatic Position Inference, Location, and Tracking (Auto-PILOT)

Visual fix sources can be acquired from any optical sensor at any time. Modern submarines are equipped with numerous sensors as part of the Integrated Submarine Imaging System (ISIS), which includes periscope video in the visible (electro-optical, EO) and various infrared (IR) spectra. However, this data is currently not utilized for navigation due to a lack of appropriate tools, yet computer vision and digital image processing technologies are capable of analyzing imagery to identify and track different types of navigation aids. In addition, digital navigation products, such as Digital Nautical Charts (DNCs) and Electronic Navigation Charts (ENCs), are available to provide relevant information about shipping lanes, navigational aids, depth soundings, and underwater hazards. This proposal describes a system that incorporates these different sources of information and provides automated tools to assist navigation and piloting teams. In particular, we will develop computer vision algorithms to provide enhanced capabilities for the automatic detection, tracking, and re-acquisition of visual location fixes. Our approach builds on the Surveillance, Persistent Observation, and Target Recognition (SPOTR) system, a suite of real-time image and video analytics tools successfully deployed with numerous government and commercial customers.